California Sea Cliff Metrics: Mapping and Validation

Abstract:

Seacliff erosion is a serious hazard with implications for coastal management, and is often estimated using successive hand digitized cliff tops or bases (toe) to assess cliff retreat. We developed an automated procedure to extract the location of the cliff top from high resolution lidar-derived digital elevation models using transects generated at approximately 1-m intervals. The automated method to define cliff tops is repeatable, takes advantage of detailed topographic information within high-resolution elevation data, and is more efficient than hand-digitizing. To validate the results obtained from a 2010 aerial lidar survey, we conducted a terrestrial lidar (tlidar) survey at Bonny Doon beach near Santa Cruz California in 2014 and mapped the location of the cliff tops using real-time kinematic GPS. Bonny Doon beach has highly irregular cliffs, with several small sea caves and erosion features that were not distinguishable in the aerial lidar digital terrain model (DTM). We extracted the location of the cliff top from the tlidar point cloud along the same transects used previously to automatically derive cliff tops from the aerial lidar derived DTM. The minimum horizontal distance between the tlidar derived cliff top points and GPS points was calculated. The error measurements between GPS and terrestrial lidar are 0.19 m mean absolute error (MAE) and 0.51 m root mean square error (RMSE) respectively. The MAE and the RMSE between the GPS and aerial lidar cliff top points were 0.96 and 1.82 m respectively. The larger errors between aerial lidar cliff top points and GPS are likely due to the failure to remove all vegetation from the aerial lidar data, the positional accuracy of the aerial lidar, and a 4-year gap between surveys. The error assessment indicates that the automated procedure mapped the cliff top location successfully using both the aerial and terrestrial lidar, although the cliff top delineation based on the tlidar data was more accurate.